In preparation for the race that stops a nation, IT systems have to saddle up as efficiently as jockeys do.
So during the Spring Racing Carnival and in the lead up to the Melbourne Cup, Centrebet CIO Stefan Matthews and his team start their days at 6:30am. Since 2005, the online betting agency's transactions have almost doubled every year.
“Looking at my wall I’ve got graphs, charts and metrics showing trends and analysis and cross referencing between not only the Melbourne Cup, but between all the core racing events during the spring carnival,” Matthews said.
“We know, from a transactional standpoint, how each event in the Spring Racing Carnival correlates with each other and how that is correlating with previous years.”
On a weekly basis, the predictive modelling is cross referenced to see how it compares with the actual results coming through Centrebet's system on an event by event basis.
Matthews said the preparation and rigourous data analysis enables his team to predict the website traffic it will encounter on Cup day.
“It gives us a stable base that we can use for monitoring and making decisions about what’s occurring in the systems,” Matthews said.
As horse racing is one of Centrebet’s core products, the Spring Racing Carnival is treated like a project. A team is set up three months prior to the event, which include members from different disciplines within the technology group.
From a network and infrastructure standpoint, the IT team does weekly tuning of system components – application servers, web servers, network infrastructure, measuring its response time in terms of milliseconds.
After each event, Matthews and his team monitor what’s happening at the database level at the web server level, looking for any improvements it can bring to the configuration of systems at that level.
“That level of granularity actually pays off when you get into the really high transactional activity that’s going to occur on Melbourne Cup day,” Matthews said. “Because as the system becomes under extreme load, that’s when you’ll see those sorts of bottlenecks start to occur.”
The project team benefits from the analytical capabilities within its SAS data warehouse, which stores 3 to 5 terabytes of data.
The online rush on Melbourne Cup day begins at 7am, reaching a low at 11am and then up until the starter pistol fires, the system is under a constant extreme load.
“When the jump occurs at 3pm my team will all be sitting at their desks and at that stage we go ‘thank God for that’!” Matthews said.
"For us the biggest load of all on our system’s infrastructure occurs once the race is completed, because at that point in time we’ve accumulated and amassed all the betting transactions that have occurred over the weeks and months."
Once the horses grace the winner's podium, transactions have to be resulted and the bets settled for the lucky punters. From a systems perspective, Matthews said it’s the peak load and it's critical to make sure the mountains of transactions are resulted quickly and efficiently.
“Over the last three years we’ve been very accurate, and we’re quietly confident that we know how much web site traffic we’re going to be looking at this cup day.”
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.